I finally found time to return to what I started in December.
..........................................................................
Do you remember our thread about supercaps? One of the issues was an idea
that supercaps are partially rechargeable batteries. How can I demostrate
that a part of the delivered energy is stored in a chemical form? I am not
a chemist and I can not compare chemical compositions of the electrolite
before and after charging (two destructive tests). But the data below can
be used to support the "chemical energy" hypothesis.
1) A supercapacitor (nominally 1 farad and 5.5 volts) was discharged by
connecting its terminals for 16 hours (to play it safe). Then it was
charged from a 4.5 source (three D batteries) through a resistor of
1000 ohms (+/- 1%). The potential difference across the resistor was
measured with a high impedance probe (every minute for 4 hours) and used
to calculate the current. The abbreviated data are shown below. Plot them
as a semi-log graph and you will notice that the decrease of current is
not exponential (as it was for several ordinary electrolytic capacitors,
such as 1000 microF, 25 volts from Nippon, and for the non-electrolytics).
Here is what I did with these data. First the initial slope (log of the
current-versus-t) was used to find C1. The result, 1.05 F, was nearly the
same as the nominal C. The total charge delivered would normally be expected
to be Q1=C1*V=1.05*4.5=4.72 coulombs. But that is not what I found from the
area under the linear current-versus-time plot. The total charge delivered
turned out to be Q2=6.09 coulombs. That is 46% higher than expected. I
infer that about 30% of the charge received (and energy) is storred in
"some unusual", probably chemical, form.
Can somebody who has 5 and 50 farads supercaps test them in the same way.
Please share the results. I suspect that the ratio of Q2/Q1 will go up
with C (the battery effect becomes progressively more significant).
2) Discharging (same C, same R but no battery). The dirrection of the
current is reversed, naturally.
t (min) 0 5 10 20 40 80 160 200
I (mA) 4.40 3.26 2.50 1.62 0.64 0.16 0.05 0.01
The initial I-versut-t is essentially the same as in charging but the tail
is slightly steeper. This means that C1 and Q1 are as before but Q2 is
smaller. In fact Q2 is 5.46 farads. This would be about 5.5 farads if an
additional step (an extrapolation) was allowed for to get I=0.00 at t=240.
I stopped earlier because I wanted to verify something I noticed in December.
Observe that the current of 0.01 mA implies that the capacitor is nearly
totally discharged (the remaining voltage is 10 mV). I disconnected the
capacitor and left it alone for four hours. Then I connected it again to
finish the discharging. The initial current was found to be 0.12 mA and
it was decreasing as shown below.
t (min) 0 3 10 20 30
I (mA) 0.12 0.09 0.05 0.01 0.00
The disconnected capacitor was somehow able to increase its voltage by
the factor of 12. The area under this plot is 0.07 farads. I suspect that
some chemical reactions were taking place inside the supercap (turning
molecular energy into electric energy). Will more charge appear on the
disconnected capacitor in the next several hours. This is possible; the
charge removed so far was only 5.46+0.07=5.53 farads. This is less than
the 6.09 farads delivered to the capacitor in charging it. I will continue
this experiment tomorrow and will share my observations. All comments will
be appreciated.
Ludwik Kowalski